effect of metakaolin on the chloride ingress properties of ... · pdf fileeffect of metakaolin...
TRANSCRIPT
KSCE Journal of Civil Engineering (0000) 00(0):1-10
Copyright 2015 Korean Society of Civil Engineers
DOI
− 1 −
pISSN 1226-7988, eISSN 1976-3808
www.springer.com/12205
Structural Engineering
Effect of Metakaolin on the Chloride Ingress Properties of Concrete
R. M. Ferreira*, J. P. Castro-Gomes**, P. Costa***, and R. Malheiro****
Received March 11, 2014/Revised November 26, 2014/Accepted June 8, 2015/Published Online
··································································································································································································································
Abstract
Research and practical experience have shown that partial replacement of cement by metakaolin improves concrete durability as aresult of the refinement of the pore structure. While much research has been presented on concrete performance with metakaolin, it isscarce concerning the transport properties of chlorides in concrete with metakaolin, in natural conditions (i.e., un-accelerated). Thisstudy determines the chloride diffusion coefficients for several concretes with vary levels of cement replacement with metakaolin,and compares these results with chloride migration coefficients obtained from accelerated laboratory testing. In this study, twocement types (CEM I 42,5R and CEM IV/A 42,5) and two cement contents levels where used with metakaolin replacement levelsvarying from 10-20%. Concretes where tested for fresh properties, compressive strength, electrical resistivity, chloride ingresscharacteristics (natural diffusion and migration), and mercury intrusion porosimetry. The results show improved strength, durabilityproperties and chloride penetration resistance of concretes with metakaolin. Furthermore, the use of metakaolin in fly ash concreteimproves the early age performance of the concrete (<90 days), counteracting the delay in strength and durability gain typicallyassociated with fly ash concrete. The results obtained from this study fulfil the lack of critical input for service life design models ofreinforced concrete structures in chloride environments, with emphasis on concretes with metakaolin replacement.
Keywords: concrete, durability, metakaolin, chloride, diffusion, migration, electrical resistivity, mercury intrusion porosimetry
··································································································································································································································
1. Introduction
Extensive research and practical experience have shown that
partial replacement of cement by metakaolin improves concrete
strength and durability as a result of the refinement of the pore
structure. Calcium Hydroxide (CH) itself does not contribute
significantly to the strength development, but its presence can
compromise the durability of concrete (Siddique and Klaus,
2009). The removal of CH coarsens the pore structure and
influences the transport mechanisms of moisture and aggressive
substances as well as deterioration mechanisms of concrete
(Sabir et al., 2001; Bai and Gailius, 2009). Metakaolin in the
concrete mix reacts with cement hydrates, CH, resulting in
calcium silica hydrates which are more stable.
There are also ecological benefits, which make concrete
produced with metakaolin a more sustainable alternative to OPC
(ordinary Portland cement) concrete. Metakaolin is produced
from calcining kaolin clay at a specific temperature range (500-
900ºC) to make it reactive, with the general form Al2O3-SiO2
(Siddique and Klaus, 2009; Sabir et al., 2001; Huat, 2006;
Ramezanianpour et al., 2012; Antoni et al., 2012; Rashad, 2013;
Srivastava et al., 2012).
The production of metakaolin releases less CO2 than OPC
which results from the de-carbonation of limestone. The lower
temperatures required to produce metakaolin mean lower energy
requirement, resulting in lower CO2 production from the fuel. As
a result, metakaolin is considered a favourable replacement for
cement due to its environmental benefit and the growing
economic and social constraints on cement production.
While much research on the mechanical properties and durability
performance of concrete with metakaolin has been presented, it
is scarce with regards to the quantification of chloride ingress in
natural conditions. Research has focused on the use of accelerated
laboratory testing procedures to quantify chloride ingress resistance,
mainly due to the time consuming process that is natural chloride
diffusion. Kim et al. (2007) show identical performance of
concrete with metakaolin or with silica fume measuring the total
electrical charge passed as an indication of the concretes ability
to resist chloride ion penetration (ASTM C 1202, 2010). Using
the same test method, Poon et al. (2006) show that the total
electrical charge passed decreases for increasing metakaolin
contents and for various water/binder ratios. Studies by Boddy et
al. (2001) show that circa 8% metakaolin in concrete can reduce
diffusion coefficients by 50%, where as 12% metakaolin can
TECHNICAL NOTE
*Senior Researcher, VTT Technical Research Centre of Finland, Espoo, Finland; C-TAC – University of Minho, Guimarães, Portugal (Corresponding
Author, E-mail: [email protected])
**Full Professor, C-MADE – University of Beira Interior, Covilhã, Portugal (E-mail: [email protected])
***Researcher, C-TAC – University of Minho, Guimarães, Portugal (E-mail: [email protected])
****Researcher, C-TAC – University of Minho, Guimarães, Portugal (E-mail: [email protected])
R. M. Ferreira, J. P. Castro-Gomes, P. Costa, and R. Malheiro
− 2 − KSCE Journal of Civil Engineering
reduce apparent diffusion coefficient (DA) by 30%. Trejo and
Halmen (2006) measured the total electrical charge passed and
DA, for concrete with varying w/b ratio and for metakaolin with
two finesses. The results show significant improvement in
performance, with decreasing w/b ratio. The effect of metakaolin
finess was not distinguishable. Gruber et al. also obtained large
decreases (50-60%) in DA with 8-12% cement replacement with
metakaolin (Gruber et al., 2001). Zeljkovic et al. (2009) also
studied the total charge passed and DA with similar results.
Nokken et al. (2006) showed similar tendencies, with 8% metakaolin
having identical performance as that of 4% silica fume. Justice et
al. (2005) measured a low rapid chloride permeability (AASHTO
T 277, 1993) for concrete with 8% cement replacement for w/b
ratios of 0.40 and 0.50. Courard et al. (2003) measured diffusion
coefficients of mortars, obtaining lower values with increasing
metakaolin contents.
Jalali et al. (2006) and Camões and Reis (2011) studied durability
properties of concrete with cement replacement with mixes of
metakaolin and fly ash. Concretes had a w/b ratio of 0.5 and 0.55
respectively. Cement replacement was 15% and 10% respectively. In
both these studies, 28 and 90 day Rapid Chloride Migration
(RCM) values are at least three times smaller than reference
values (CEM II). The performance of concrete with both
metakaolin and fly ash is even better than that of just metakaolin.
Camões et al. (2004a) and Camões et al. (2004b) also studied
the use of metakaolin and latex for bridge concrete. RCM and
total charge passed test were performed on concrete with 10%
replacement of metakaolin and w/b ratio of 0.45. The results
showed a decrease in the test value by a factor of 3 for both tests
in relation to the reference concrete (CEM I).
Coleman and Page (1997) showed that pastes with metakaolin
(10-20%) had higher binding capacities. Despite the reduction in
pore solution hydroxide ion concentration from long-term
observations, and the reduction in pore solution pH, since the
[Cl−]/[OH−] ratios are similar to those of Portland cement pastes,
Coleman and Page concluded that the inclusion of up to 20%
metakaolin would have little effect on the risks for chloride
induced corrosion of reinforcing steel.
Aguayo et al. (2014) tested concretes with both metakaolin
and limestone replacement of cement. Metakaolin replacement
level was 10% and for limestone 10% and 25%. Rapid chloride
permeability (charge passed) and non-steady state migration tests
were performed. The results show the beneficial impact of
metakaolin on concrete performance, even with levels of up to
25% limestone when compared to OPC concrete. Migration
coefficients for the OPC concrete were approximately 2-3 times
greater than those of the concrete with metakaolin and limestone
replacement.
The literature review presented reveals that chloride ingress
has been researched mainly by using accelerated test methods
(chloride migration testing), and by indirect test methods that
measure parameters that are well correlated, but do not define
adequately, the ingress of chlorides (electrical resistivity, current
passed).
Therefore, the main focus of this study is to analyse the
chloride ingress properties of these concrete, using both natural
diffusion and accelerated migration methods, and quantify the
performance so that durability parameters needed for performance
based approach are defined. The results obtained from this study
fulfil the need of critical input for service life design models for
reinforced concrete structures in chloride environments.
2. Experimental Studies
Three concrete mixes are used in this study. Two mixes use a
Portland cement (CEM I 42,5 R) with different cement contents
(330 kg/m3 and 440 kg/m3), and a third mix with a fly ash cement
(CEM IV/A 42,5) with a cement content of 440 kg/m3. All mixes
are prepared with varying replacement ratios of metakaolin (0-
20% by weight of cement). In the following, the concrete mixes
are referred to by their cement content level and percentage of
metakaolin replacement, i.e., 300-10 refers to the concrete mix
with 330 kg/m3 of Portland cement in which 10% has been
replaced with metakaolin; 440FA-15 refers to the concrete mix
with 440 kg/m3 of fly ash cement in which 15% has been
replaced with metakaolin. In addition to the fresh properties of
the concretes mixes, the following hardened properties were also
measured: compressive strength, electrical resistivity (4-point
Wenner electrode), chloride diffusion, chloride migration, and
Mercury Intrusion Porosimetry (MIP).
2.1 Materials
Two cements were used in the production of concrete
specimens: CEM I 42,5R and CEM IV/A (V) 42,5 in accordance
to the NP EN 197-1 (2001). A commercially available metakaolin
was used - Class N according to the ASTM C 618 (2008). In
Table 1 the chemical and physical characteristics of the cements
and metakaolin used are presented.
Two crushed granite coarse aggregates and a river sand were
used. The coarse aggregates with sizes 5-15 mm and 15-30 mm
have specific gravities of 2.63 g/cm3 and 2.67 g/cm3, respectively.
The water absorption is 1.60% and 1.38%, respectively. The
Table 1. Properties of Cements and Metakaolin
Parameters CEM I 42,5 RCEM IV/A (V) 42,5
Metakaolin(MK)
SiO2 (%) 19.55 38.06 51.50
Al2O3 (%) 4.24 4.67 44.50
K2O (%) - - 0.21
Na2O (%) - - 0.11
Fe2O3 (%) 3.34 3.80 0.45
CaO (%) 62.61 43.97 0.02
SO3 (%) 3.26 3.00 -
MgO (%) 2.51 2.51 0.12
Cl− (%) 0.03 0.03 -
Loss on ignition (%) 2.96 2.53 < 0.80
Specific gravity 3.12 2.84 2.20
Specific Surface (m2/g) 0.41 0.44 15-18
Effect of Metakaolin on the Chloride Ingress Properties of Concrete
Vol. 00, No. 0 / 000 0000 − 3 −
river sand had a specific gravities of 2.63 g/cm3 and an absorption
of 1.04%.
2.2 Mix Design
Mix design was based on maintaining a constant water/binder
(w/b) ratio. The 440 and 440FA concrete mixes have a 0.45 w/b
ratio, and the 330 concrete mix has a 0.60 w/b ratio. Each mix
series was comprised of a references mixture with no metakaolin,
and three additional mixtures with 10%, 15% and 20% of cement
replacement with metakaolin. Mix 440FA was not made for 20%
replacement of metakaolin. Details of the concrete mix designs
and their workability are presented in Table 2. Concrete
workability was kept within the S2/S3 class of the NP EN 206-1
(2007), see slump values in Table 2.
The concretes were produce in a vertical axis mixer with a 120
litre capacity. Cubic specimens of 100 mm and 150 mm in
dimension and cylinders of 100 mm in diameter and 200 mm in
height were cast in steel moulds and compacted using a vibrating
table. 24 hours after casting the specimens were removed from
the moulds and permanently submerged in water tanks at 20±3ºC
until testing. Testing was performed at 7, 14, 28, 90 and 180
days.
2.3 Compressive Strength
The compressive strength of the concrete was evaluated using
150 mm cubes. The tests were performed on an ELE compressive
machine with 3000 kN capacity, at a loading rate of 0.3 N/s. Each
result represents an average of 3 specimens. The compressive
strength tests were performed according to the NP EN 12390-3
(2009).
2.4 Electrical Resistivity
The electrical resistivity of the concrete was determined using
a four-point electrode resistivity-meter (RM MK II - alternating
trapezoidal current wave with frequency of 13 Hz) on 150 mm
cubic concrete specimens. Measurements were performed with
40 mm electrode spacing. The apparent electrical resistivity
determined by the equipment is obtained from Eq. (1) considered
valid for a homogeneous semi-finite volume of the material.
(1)
Where ρA is the electrical resistivity (Ωm), a the electrode
spacing (m), V the voltage drop (V), and I the current (A).
According to LMC testing procedure for electrical resistivity of
concrete (PE-002, 2005), cubic specimens are measured on three
adjacent surfaces with two measurements per surface perpendicular
to each other.
2.5 Non-steady State Chloride Migration
The non-steady state chloride migration coefficient (DNSS) was
determined on cylindrical specimens of 100 mm diameter and 50
mm height according to the NT BUILD 492 (1999). Each result
represents an average of 3 specimens. The applied electrical
potential and duration of testing are defined according to the
initial current measured across the concrete specimen for a
reference electrical potential of 30 V. After testing, specimens are
split in half and silver nitrate solution is sprayed onto the newly
exposed surface. Based on the measurement of chloride penetration
depth, the non-steady state chloride migration coefficient is
determined.
2.6 Chloride Immersion
The chloride immersion is a natural diffusion test performed
on 100 mm cubic specimens according to the LNEC E390
(1993). Concrete specimens are immersed in a 3% NaCl solution
for 90 days after which powder samples are taken at 5 mm
interval in depth from the surface. The chloride profiles are
determined by chloride titration of the powdered samples. Each
result represents an average of 3 specimens. Curve fitting
chloride profiles with the solution of Fick’s 2nd Law of diffusion
results in the determination of the apparent diffusion coefficient
(DA) and the surface concentration of chlorides (CS).
2.7 Mercury Intrusion Porosimetry
The porosimetry testing was performed on concrete samples
90 days of age, with an Auto Pore IV scanning mercury porosimeter
ρA 2πaV I⁄=
Table 2. Concrete Compositions and Workability
Series w/bWater(l/m3)
Binder (kg/m3) Aggregate (kg/m3)Slump (mm)
Cement MK Coarse 1 Coarse 2 Sand
330-0
0.60
198 330 0
405 835 710
90
330-10 198 297 33 130
330-15 198 280.5 49.5 140
330-20 198 264 66 145
440-0
0.45
200 440 0
420 820 600
90
440-10 200 396 44 120
440-15 200 374 66 125
440-20 200 352 88 130
440FA-0
0.45
200 440 0
420 820 600
95
440FA -10 200 396 44 90
440FA-15 200 374 66 85
R. M. Ferreira, J. P. Castro-Gomes, P. Costa, and R. Malheiro
− 4 − KSCE Journal of Civil Engineering
having a pressure range from sub-ambient to 33,000 psi. The
contact angle and the surface tension of mercury considered
were 130° and 0.485 N/m (Abell et al., 1999), respectively. Prior
to testing, specimens were oven-dried at 105°C for 24 hours. To
obtain representative pore-size distribution curve for the concrete
samples, the results of two porosimetry data were averaged.
The mercury porosimetry graphs showing the distribution of
pore diameters are givem. Along the y-axis the log differential
intrusion in ml/g, or dV/dlogD, where V is the volume of mercury
intruded into the pores of the sample. This is the derivative of
intruded volume with respect to the logarithm of pore diameter.
The derivative with respect to log D is used instead of the
derivative with respect to D to amplify the dV/dD values. Thus,
the derivative plot clearly identifies points of inflection, which
shows where clusters of pores of a particular diameter occur.
3. Results and Discussion
3.1 Compressive Strength
The results of the compressive strength tests (average and
coefficient of variation) are presented in Table 3.
The 28 day compressive strength values show that, for all
concrete, there is no strength loss when metakaolin is used in the
mixes.
For the 330 concretes, due to the higher water/binder content
the effect of metakaolin on strength development is noticeable
after 28 days. At 14 days, only the 10% and 15% metakaolin
replacement match the reference concrete. From 28 days
onwards, the compressive strength of concrete with metakaolin
is identical or higher strength than that of the reference concrete.
For the 440 concretes, at 7 days the 10% and 15% metakaolin
replacement match the reference concrete. By 14 days, all
concrete have identical or greater compressive strength than the
reference concrete. At 180 days, concrete 440-15 has a 10 MPa
greater compressive strength than 440-0 (20% increase in
strength).
For the 440FA concrete, due to the fly ash in the mix, the
strength development is known to be delayed. The results show
that the metakaolin compensates the use of fly ash by providing
the initial early age strength gain. At 7 days the 10% metakaolin
replacement shows a gain of 9% in relation to the reference
whereas the 15% metakaolin replacement shows a loss of 8% in
relation to the reference. At 14 days, both replacement percentages
show gains in strength in relation to the reference concrete. By
180 days, these gains are of 11% and 15% for the 10% and 15%
metakaolin replacement, respectively.
In summary, the 7 day compressive strength of the 440 and
440FA concretes with metakaolin replacement match that of the
reference concrete (considering statistical variability), with the
exceptions of 440-20 and 440FA-15 which represent the mixes
with the highest percentage of metakaolin. By 14 days, all
concretes with metakaolin replacement match or exceed the
compressive strength of the reference concretes. For the 330
concrete, this tendency is observed only after 28 days.
3.2 Electrical Resistivity
In Table 3 the results for electrical resistivity measurements are
presented.
For the 330 and 440 concretes, 7 day electrical resistivity
values are generally lower than those of the reference mixes, but
by 28 days the tendency is reversed. The concretes with metakaolin
replacements show an increase between 47-69% in electrical
Table 3. Compressive Strength and Electrical Resisitivty Results
SeriesAverage compressive strength (MPa) (CoV) Average electrical resisitivty (Ωm) (CoV)
7 days 14 days 28 days 90 days 180 days 7 days 28 days 90 days 180 days
330-018.9
(1.70)22.2
(1.71)23.9
(2.15)28.1
(3.92)29.1
(0.79)30.0
(4.11)39.5
(4.64)49.5
(5.76)50.8
(5.12)
330-1017.0
(1.18)22.2
(0.78)26.6
(0.38)31.1
(5.58)31.8
(1.73)25.6
(5.54)73.1
(4.22)74.1
(6.13)74.9
(5.46)
330-1514.9
(1.55)21.0
(2.20)25.0
(2.01)31.5
(1.20)33.1
(1.98)21.7
(7.07)65.4
(5.99)65.7
(6.67)66.4
(7.58)
330-2012.4
(0.93)18.8
(0.62)23.4
(2.38)29.6
(2.91)33.6
(0.52)20.3
(6.71)62.5
(4.82)63.2
(4.68)65.9
(5.04)
440-030.5
(6.27)33.7
(4.75)40.6
(1.37)44.0
(0.35)44.8
(0.68)38.3
(5.91)54.8
(4.80)71.6
(5.82)72.5
(5.97)
440-1030.2
(4.80)35.9
(0.85)40.9
(1.29)44.8
(2.72)45.4
(2.20)32.6
(5.34)95.5
(4.53)95.9
(4.50)96.7
(7.57)
440-1532.7
(2.64)40.4
(4.49)48.0
(0.94)53.7
(2.70)54.6
(2.21)39.3
(5.34)124.8(3.81)
127.1(3.07)
128.0(3.76)
440-2024.7
(1.24)32.1
(3.76)39.7
(0.63)46.8
(1.82)47.6
(1.49)30.2
(5.72)86.0
(5.93)107.6(3.75)
112.5(4.23)
440FA-022.4
(5.32)27.1
(3.92)29.7
(2.91)38.5
(2.02)43.3
(2.02)34.5
(5.27)58.5
(4.74)98.8
(5.10)205.0(4.80)
440FA -1024.5
(5.44)32.6
(3.95)37.9
(1.99)45.6
(2.64)48.0
(0.52)50.1
(5.32)185.2(4.12)
227.0(5.56)
276.8(4.96)
440FA-1520.6
(4.23)30.5
(3.84)36.4
(0.69)43.5
(2.74)50.3
(0.46)35.2
(4.77)137.9(5.11)
184.9(5.59)
248.7(5.78)
Effect of Metakaolin on the Chloride Ingress Properties of Concrete
Vol. 00, No. 0 / 000 0000 − 5 −
resistivity values due to the refinement of the pore structure as a
result of the pozzolanic reaction. For 330 concrete, the 10%
metakaolin replacement level has the highest electrical resistivity
value at 180 days (74.9 Ωm−30% greater than 330-0). For the
440 concrete, the 15% metakaolin replacement level reveals the
highest electrical resistivity values at 180 days (128.0 Ωm−43%
greater than 440-0).
For the 440FA concrete, already at 7 days the electrical resistivity
values of the concretes with metakaolin replacement are between
2-31% higher than the reference concrete. At 180 days, both the
reference and the concretes with metakaolin replacement show
high electrical resistivity. This is due to the delayed hydration effect
of fly ash. The concrete with metakaolin replacement are, however,
17-25% higher. The 440FA-10 has the highest electrical resistivity
value at 180 days (276.8 Ωm−25% greater than 440FA-0).
Given the good correlation electrical resistivity has with other
durability indicators (Ferreira, 2010; Hornbostel et al., 2013;
Ferreira, 2000; Khater, 2011) this reflects the beneficial effect of
metakaolin replacement.
3.3 Non-steady State Chloride Migration
In Table 4 the chloride migration coefficient is presented
(average of 3 specimens).
For each of the three concrete mixes, 7 day values are identical
(taking into consideration the scatter) meaning the influence of
metakaolin on the chloride migration coefficient is not observed.
By 28 days, the effect of metakaolin is noticeable.
For the 330 concrete, by 28 days the concrete mixes with
metakaolin have lower migration coefficients than the reference
mix (330-0). However, from 28 days onwards there is no
significant improvement in the migration coefficients. By 180
days, for all 330 concrete with percentages of metakaolin
replacements, the migration coefficient is roughly half that of the
reference mix (DNSS, 330-0 = 29.4 10−12 m2/s).
For the 440 concrete, by 180 days the reference concrete’s
migration coefficient has reduced to approximately half. For the
concretes with metakaolin replacements, the reduction in time is
greater than that of the reference concrete. Comparatively to the
180 day reference concrete migration coefficient, the reduction is
between 37-56% for the concrete with 10% and 15% metakaolin
replacement, respectively.
For the 440FA concrete, the presence of the fly ash is reflected
in the delayed improvement in the properties of the concrete.
While at 7 days the chloride migration coefficient of 440FA-0 is
only marginally lower than the mixes with metakaolin, by 90
days it is double the value of the mixes with metakaolin. This
shows how the metakaolin has improved the migration coefficient
of the fly ash concrete at early ages. However, by 180 days, the
reference concrete 440FA-0 has an almost identical migration
coefficient. Here, the larger amount of fly ash in the concrete and
the longer hydration time has helped it recover to the level of the
mixes with metakaolin. These results show that metakaolin can
help accelerate the gain of the migration coefficient without
hinder the long term values. It is at the early ages when concrete
is most vulnerable to penetration of chlorides that the rapid
improvement of concrete resistance to chloride ingress is desired.
3.4 Chloride Immersion – Natural Diffusion
The chloride profiles measured from the specimens subject to
the immersion test are presented in Figs. 3-6. In Fig. 3, the
Table 4. Results of the Non-steady State Migration and Immersion Test
Series
Non-steady state chloride migration coefficient DNSS (10−12 m2/s)
Apparent diffusion coefficient DA (10−12 m2/s) / Surface chloride concentration CS (%/mass binder)
7 days 28 days 90 days 180 days DA CS Depth of 0.6% Cl
330-054.0
(6.05)31.3
(2.69)30.7
(3.74)29.4
(4.00)12.00 4.2 ≈ 21
330-1047.8
(7.01)15.8
(2.35)13.4
(6.49)13.4
(9.74)3.52 4.0 ≈ 12
330-1559.1
(4.04)17.1
(9.17)14.3
(11.01)14.2
(4.15)3.95 5.0 ≈ 10
330-2051.1
(4.91)23.1
(4.09)15.2
(2.89)14.9
(3.07)2.78 2.9 ≈ 8
440-025.1
(4.50)15.5
(3.02)13.2
(4.98)13.2
(7.75)4.82 2.0 ≈ 12
440-1026.7
(8.16)8.5
(3.17)8.3
(5.93)8.1
(2.48)3.62 2.3 ≈ 10
440-1524.2
(1.23)7.0
(10.63)6.6
(9.93)5.8
(12.18)2.32 2.8 ≈ 7
440-2027.4
(8.78)10.5
(2.91)7.3
(3.84)6.3
(9.11)1.35 3.0 ≈ 7
440FA-021.8
(3.24)21.5
(5.63)11.4
(8.02)5.7
(4.44)5.27 2.8 ≈ 12
440FA -1023.7
(6.33)7.8
(2.15)4.6
(13.30)4.5
(4.00)1.97 3.7 ≈ 7
440FA-1522.6
(2.04)10.3
(18.25)5.8
(19.62)5.4
(8.76)1.82 3.7 ≈ 7
R. M. Ferreira, J. P. Castro-Gomes, P. Costa, and R. Malheiro
− 6 − KSCE Journal of Civil Engineering
chloride profiles of all three reference mixes are presented
together. The results of the curve fitting of Fick’s 2nd Law of
diffusion to the chloride profiles (DA and Cs) are presented in
Table 4. These results are for concretes that have been exposed to
chlorides at the age of 28 days, and have been immersed in water
for 90 days.
For the 330 concrete, the effect of metakaolin replacement is
observed by a large reduction (≈75%) in the apparent diffusion
coefficient compared to the value of the reference concrete 330-
0. This implies a significant improvement in the performance of
the concretes with metakaolin, even though a high water/binder
ration is used. This result shows that the migration testing,
despite showing similar general tendencies, is not capable of
reflecting the same differences in the qualities of the concrete.
For the 440 concrete, produced with a low water/binder ratio,
the difference between the reference 440-0 and the mixes with
metakaolin is reduced. The results show that the improvement in
the apparent diffusion coefficient is proportional to the amount of
metakaolin used. This means the smallest improvement is
observed for the 440-10 concrete (25%−3.6 10−12 m2/s), and the
greatest improvement is seen for the 440-20 concrete (70%−1.4
10−12 m2/s).
For the 440FA concrete follows the same tendency as observed
for the 440 concrete. This is because the 440FA concrete tested
did not have enough time to hydrate and benefit from the
reaction of the fly ash. However, as mentioned previously
regarding the migration coefficient, the impact of metakaolin on
the early age apparent diffusion coefficient is beneficial. For both
the 440FA-10 and 440FA-15 concretes a reduction to approximately
a third of the apparent diffusion coefficient of the reference
concrete is observed (2.0 10−12 m2/s, and 1.9 10−12 m2/s, respectively).
The natural diffusion test has an advantage over accelerated
method (especially electrically induced) in that it allows for the
binding reaction of the chlorides to the cement hydrates.
Furthermore, concretes with metakaolin replacement show
higher surface chloride concentrations (associated with higher
binding capacity), and have smaller volume of chloride penetration,
especially in the intermediate depths (8-20 mm) as a result of the
finer pore structure developed. This effect is neither measurable
nor taken into account in the accelerated migration test.
3.5 Mercury Intrusion Porosimetry
Figure 1 shows mercury intrusion porosimetry graphs for the
reference concrete mixes 330-0 and 440-0. The pore-size
distribution curves are of similar shape as would be expected.
Both curves show a cluster of pore-size diameters in the range of
0.02 to 0.5 microns. However, the porosity data for 330-0
concrete give a total intrusion volume of 0.11 mL/g (or 1.1 ×
10−7 m3/g), while the total intrusion volume for 440-0 concrete
is only 0.05 mL/g (or 0.5 × 10−7 m3/g). So, total intrusion
volume of 440-0 is about half of 330-0 but both showing a
cluster of pore-size diameters around 0,1 microns (from 0.02 to
0.5 microns).
In Fig. 2 mercury porosimetry graphs are presented comparing
330-0 reference concrete and concrete mixes with 10%, 15% and
20% metakaolin replacement. It shows that total intrusion
volume for concrete with 15% and 20% metakaolin replacement
is slightly higher than reference concrete, except for 10%
metakaolin replacement. However, the pore-size diameters of
concrete with metakaolin replacement are in the range of 0.01 to
0.2 microns, which clearly confirms the refinement of concrete
structure; ie., the porosity within range of 0.2 to 0.5 microns has
been annulled while small size pores (from 0.01 to 0.02 microns)
have increased.
Similar results, although not so obvious, were also found for
the 440 FA concrete mixes and having metakaolin replacement.
The total pore volume was found to increase with increase in
metakaolin replacement. Similar influences by metakaolin on the
pore structure and diffusion rates of metakaolin-Portland cement
pastes have been reported (Frias and Cabrera, 2000; Cabrera and
Nwaubani, 1998; Khatib and Wild, 1996).
Fig. 1. Chloride Profiles for the Reference Concretes Without Metaka-
olin
Fig. 2. Chloride Profiles for the 330 Concrete Mixes
Effect of Metakaolin on the Chloride Ingress Properties of Concrete
Vol. 00, No. 0 / 000 0000 − 7 −
4. Quantifying the Benefit of Metakaolin Replace-ment on the Durability Performance of Concretein Chloride Laden Environment
The durability performance of concrete with metakaolin
replacements improves not only due to the increase in resistance
to chloride ingress (i.e., improvement in the diffusion coefficient),
but also due to the rate with which this decrease takes place in
time. A calculation estimating the service life of a reinforced
concrete structure in chloride environment has been used to
demonstrate the performance of concrete mixes studied (Ferreira,
2010b; Ferreira, 2008). The model adopted for chloride penetration
of concrete is described in detail in the fib Model Code for
Service Life Design – Bulletin 34 (2006).
A serviceability limit state of corrosion initiation was considered
for this service life calculation. The limit state is considered to
have failed when the chloride content at the depth of the
reinforcement has exceeded a critical chloride content. The
calculation is performed by Monte Carlo simulation with 106
simulations for each individual year. In Table 5 the parameters
Fig. 3. Chloride Profiles for the 440 Concrete Mixes
Fig. 4. Chloride Profiles for the 440FA Concrete Mixes
Fig. 5. Log Differential Intrusion Versus Pore Size for the Refer-
ence Concrete Mixes 330-0 and 440-0
Fig. 6. Log Differential Intrusion Versus Pore Size for the 330 Con-
crete Mixes
Table 5. Parameter Definition for Service Life Design Calculation
(normally distributed)
SeriesDNSS CS CCR xC n
(10-12 m2/s) (%/wt.binder) (%/wt.binder) (mm) (−)
330-0 31.3/6.0
3.0/0.5 0.6/0.05 55/5
0.4/0.04
440-0 15.5/3.1 0.4/0.04
440FA-0 21.4/4.0 0.6/0.06
330-10 15.8/1.5
3.0/0.5 0.6/0.05 55/5
0.55/0.05
440-10 8.5/1.5 0.55/0.05
440FA-10 7.8/1.6 0.65/0.65
Deterministic parameters: t0 = 28 days; T = 20oC; tE = 0.XC – concrete cover depth; DA – Apparent diffusion coefficient; CS – Surfacechloride concentration; CCR – Critical chloride concentration; and, n – Ageingfactor for DNSS
R. M. Ferreira, J. P. Castro-Gomes, P. Costa, and R. Malheiro
− 8 − KSCE Journal of Civil Engineering
and values used in the calculation are presented. For simplicity,
only normal distributions were used to describe the stochastic
variables. The service life design calculation was performed for
the reference concretes and the mixes with 10% metakaolin
replacement.
For chloride migration, laboratory values were used and the
corresponding standard deviations. The surface chloride content
was based on typical design values for marine structures along
the Portuguese coast in an XS3 environment, as was the concrete
cover (LNEC E465, 2007). Coefficients of variations were
assumed for the parameters. The choice of ageing factors for the
concrete with metakaolin replacement was difficult due to the
lack of published results. Values have been assumed based on the
expected performance in relation to the reference concrete. Less
conservative values have been chosen for the reference concrete.
Coefficients of variations of 10% were used.
In Figs. 7 and 8, the probability of exceeding the serviceability
limit state of corrosion initiation (probability of failure - pf), for a
50 year service life period is presented for the reference concrete
mixes and the mixes, with 10% replacement, respectively.
Figures 7 and 8 present the outcome of the service life
calculation. Service life calculation are meant to assist designers
in making informed decisions about the long-term durability
performance of different concrete qualities. To be able to
compare the results of the service life calculations, a criteria for
limit state acceptance has been defined, based on suggested
reliability indexes (β) for serviceability limit states (fib Bulletin
34, 2006), i.e., β = 1.3 (pf ≈ 0.1). The service life calculations
show that for all three reference concretes (330-0, 440-0, and
440FA-0), the criteria for limit state verification is already
exceeded within the first 10 years, meaning that corrosion will
start with in the first 10 years. Given that two of the concrete use
OPC this is not unexpected.
When the concrete mixes contain 10% metakaolin, the outcome
of the service life calculations show significant improvement. The
criteria for limit state verification is exceeded after 9, 35 and >50
years for concrete 330-10, 440-10 and 440FA-10, respectively.
The poor service life result for the 330-10 concrete is expected
because the concrete quality is not adequate for such an
aggressive marine environment (high w/b ratio). For the 440-10
concrete, the service life is extended approximately 30 years by
the addition of metakaolin, but further changes are required for
the design to fulfil the defined service life. This can be achieved
by either increasing concrete cover or making changes to the
concrete composition to improve its performance. In the case of
the 440FA-10, the service life is more than adequate, fulfilling
easily the criteria for limit state verification.
The service life calculation shows how the replacement of
cement with 10% metakaolin can improve the long term
performance of the concrete.
5. Conclusions
The objective of this study is to investigate the durability
performance of concrete produced with varying levels of metakaolin
replacement. The main focus is to analyse the chloride ingress
properties of these concrete, using both natural diffusion testing
(apparent diffusion coefficient) and accelerated laboratory testing
(migration coefficient). The results obtained from this study
fulfil the need for critical input for service life design models of
reinforced concrete structures in chloride environments, with the
emphasis on concretes produced with partial replacement of
cement by metakaolin. Additional characteristics of the concretes
studied have also been presented.
The main conclusions of the research are:
1. Metakaolin replacement of both Portland cement and fly ash
cement shows rapid early age improvement in the migration
coefficient (approximately half that of the reference value).
For the fly ash cement the beneficial effect of fly ash is seen
in the migration coefficient measured at 90 and 180 days.
2. Chloride apparent diffusion coefficients show a large improve-
Fig. 7. Probability of Corrosion Initiation as a Function of time for
the Reference Concretes Without Metakaolin
Fig. 8. Probability of Corrosion Initiation as a Function of Time for
the Concretes Mixes with 10% Metakaolin Replacement
Effect of Metakaolin on the Chloride Ingress Properties of Concrete
Vol. 00, No. 0 / 000 0000 − 9 −
ment in the performance of the concrete against chloride
ingress, for all percentages of metakaolin replacement, which-
ever the cement type. The scale of this improvement is
somewhat different to that observed in the migration coeffi-
cients. This can attributed to the binding of chlorides that
does not occurs in accelerated migration tests.
3. The use of metakaolin in concrete mixes with fly ash cement
improves the early age properties of the concrete counteract-
ing the delayed reaction of fly ash.
4. Metakaolin replacement (up to 20%) improves the micro-
structure of the concrete, when both Portland and fly ash
cement are used. MIP results show a slight increase in the
total porosity. However, a shift towards smaller pore sizes is
clearly visible.
5. Both compressive strength and electrical resistivity mea-
surements reflect the beneficial improvement in the pore
structure as a result of the reaction between metakaolin and
the calcium hydroxide from cement hydration. The contin-
ued improvement of electrical resistivity with time is also
observed, especially for higher metakaolin replacement lev-
els, and for concretes with lower water/binder ratios.
The results show improved strength, durability properties and
chloride penetration resistance of concretes with metakaolin.
Furthermore, the use of metakaolin in fly ash concrete improves
the early age performance of the concrete (<90 days), counteracting
the delay in strength and durability gain typically associated with
fly ash concrete. The results obtained from this study fulfil the
lack of critical input for service life design models of reinforced
concrete structures in chloride environments, with emphasis on
concretes with metakaolin replacement. In light of the growing
tendency towards concrete performance-based standards, this
study provides a needed contribution for service life design.
Acknowledgements
The authors acknowledge the support provided by the Foundation
for Science and Technology, of the Ministry of Science, Technology
and Higher Education of Portugal (PTDC/ECM/69565/2006).
References
AASHTO Standard T277-93 (1993). Standard method of test for electrical
indication of concrete's ability to resist chloride, American Association
of State Highway & Transportation Officials, Washington DC.
Abell, A. B., Willis, K. L., and Lange, D. A. (1999). “Mercury intrusion
porosimetry and image analysis of cement-based materials.” Journal
of Colloid & Interface Science, Vol. 211, No. 1, pp. 39-44, DOI:
10.1006/jcis.1998.5986.
Aguayo, M., Yang, P., Vance, K., Sant, G., and Neithalath, N. (2014)
Electrically driven chlorides ion transport in blended binder
concretes: Insights from experiments and numerical simulations,
C&CR 66, pp. 1-10.
Antoni, M., Rossen, J., Martirena, F., and Scrivener, K. (2012). “Cement
substitution by a combination of metakaolin and limestone.” Cement
& Concrete Research, Vol. 42, No. 12, pp. 1579-1589, DOI:
10.1016/j.cemconres.2012.09.006.
ASTM C 1202 (2010). Standard test method for electrical indication of
concrete's ability to resist chloride ion penetration, ASTM.
ASTM C 618 - 08a (2008). Standard specification for coal fly ash and
raw or calcined natural pozzolan for use in Concrete, ASTM.
Bai, J. and Gailius, A. (2009). “Consistency of fly ash and metakaolin
concrete.” Journal of Civil Engineering & Management, Vol. 15,
No. 2, pp. 131-135, DOI: 10.3846/1392-3730.2009.15.131-135.
Boddy, A., Hooton, R. D., and Gruber, K.A. (2001). “Long-term testing
of the chloride-penetration resistance of concrete containing high-
reactivity metakaolin.” Cement & Concrete Research, Vol. 31, No. 5,
pp. 759-765, DOI: 10.1016/S0008-8846(01)00492-6.
Cabrera, J. and Nwaubani S. (1998). The microstructure and chloride
ion diffusion characteristics of cements containing metakaolin and
fly ash, VI CANMET/ACI Int. Conf. Fly ash, Silica Fume, Slag and
Natural Pozzolans in Concrete, Vol. I. Bangkok, pp. 493-506.
Camões, A., Cruz, P., Jalali, S., and Ferreira, R. M. (2004b). Increasing
concrete durability with metakaolin and latex – A case study,
BE2004. Encontro Nacional Betão Estrutural. Porto, 17-19 November,
pp. 227-234.
Camões, A. and Reis, R. (2011). “Eco-efficient ternary mixtures
incorporating fly ash and metakaolin.” Int. Conf. Sustainability of
Constructions - Towards a better built environment. Innsbruck, 3-5
February, pp. 71-78.
Camões, C., Cruz, P., Jalali, S., Ferreira, R. M., and Cunha, P. (2004a).
Durability performance of concrete bridge piers made with metakaolin
and latex mixes, IABMAS'04. 2nd Int. Conf. Bridge Maintenance
Safety & Management. Kyoto, 18-22 October
Coleman, N. and Page, C. L. (1997). “Aspects of the pore solution chemistry
of hydrated cement pastes containing metakaolin.” Cement & Concrete
Research, Vol. 27, No. 1, pp. 147-154, DOI: 10.1016/S0008-
8846(96)00184-6.
Courard, L., Darimont, A., Schouterden, M., Ferauche, F., Willem, X.,
and Degeimbre, R. (2003). “Durability of mortars modified with
metakaolin.” Cement & Concrete Research, Vol. 33, No. 9, pp.
1473-1479, DOI: 10.1016/S0008-8846(03)00090-5.
Ferreira, R. M. (2000). Assessment of concrete durability tests, Master
Thesis, University of Minho, School of Engineering, Guimarães (In
Portuguese), p. 245.
Ferreira, R. M. (2008). “Improving durability performance of reinforced
concrete structures with probabilistic analysis.” Int. J. Concrete
Struct & Materials, Vol. 2, No. 2, pp. 137-144.
Ferreira, R. M. (2010). “Quality control of concrete with electrical
resistivity measurements.” Concrete Plant Int. – CPI, 05, pp. 82-87.
Ferreira, R. M. (2010b). “Optimization of RC structure performance in
marine environment.” Engineering Structures, Vol. 32, No. 5, pp.
1489-1494.
Frias, M. and Cabrera, J. (2000). “Pore size distribution and degree of
hydration of metakaolin - cement pastes.” Cement Concrete Research,
Vol. 30, No. 4, pp. 561-569, DOI: 10.1016/S0008-8846(00)00203-9.
Gruber, K. A., Ramlochan, T., Boddy, A., Hooton, R. D., and Thomas,
M. D. A. (2001). “Increasing concrete durability with high–reactivity
metakaolin.” Cement & Concrete Composites, Vol. 23, No. 6, pp.
479-484, DOI: 10.1016/S0958-9465(00)00097-4.
Hornbostel, K., Larsen, C. K., and Geiker, M. R. (2013). “Relationship
between concrete resistivity and corrosion rate – A literature
review.” Cement & Concrete Composites, Vol. 39, pp. 60-72, DOI:
10.1016/j.cemconcomp.2013.03.019.
Huat, O. C. (2006). Performance of concrete containing metakaolin
as cement replacement material, Master Thesis, Faculty of Civil
R. M. Ferreira, J. P. Castro-Gomes, P. Costa, and R. Malheiro
− 10 − KSCE Journal of Civil Engineering
Engineering, Universiti Teknologi Malaysia, p. 103.
Jalali, S., Peyroteo, A., J., and Ferreira, R. M. (2006). “From waste
product to metakaolin – Beneficial impact on performance parameters.”
Proc. Int. RILEM Wksp. Performance based evaluation and indicators
for concrete durability, Madrid, 19-21 March, pp. 327-333.
Justice, J. M., Kennison, L.H., Mohr, B.J., Beckwith, S.L., McCormick,
L.E., Wiggins, B., Zhang, Z. Z., and Kurtis, K. E. (2005). “Comparison
of two metakaolins and silica fume used as supplementary cementitious
materials.” 7th Int. Symp. Utilization of High-Strength/High
Performance Concrete, SP-228, Vol. 1, ACI, pp. 213-236.
Khater, H. (2011). “Influence of metakaolin on resistivity of cement
mortar to magnesium chloride solution.” J. Materials Civil Engineering,
Vol. 23, No. 9, pp. 1295-1301, DOI: 10.1061/(ASCE)MT.1943-
5533.0000294.
Khatib, J. M. and Wild, S. (1996). “Pore size distribution of metakaolin
paste.” Cement & Concrete Research, Vol. 26, No. 10, pp. 1545-
1553, DOI: 10.1016/0008-8846(96)00147-0.
Kim, H. S., Lee, S. H., and Moon, H. Y. (2007). “Strength properties and
durability aspects of HSC using Korean metakaolin.” Construction
& Building Materials, Vol. 21, No. 6, pp. 1229-1237, DOI: 10.1016/
j.conbuildmat.2006.05.007.
LNEC E 390 (1993). Concrete, Determining the resistance to chloride
penetration, Immersion test. LNEC, Lisbon, p. 2 (in Portuguese).
LNEC E 465 (2007). Concrete, Methodology for estimating the performance
properties of the concrete allowing to comply with the design working
life of the reinforced or prestressed concrete structures under the
environmental exposures XC and XS (in portuguese) LNEC, p. 24.
Model Code for Service Life Design (2006). Bulletin 34, FIB, Lausanne, p.
206.
Nokken, M. R., Boddy, A., Hooton, R. D., and Thomas, M. D. (2006).
“Time–dependant diffusion in concrete – Three laboratory studies.”
Cement & Concrete Research, Vol. 36, No. 1, pp. 200-207, DOI:
10.1016/j.cemconres.2004.03.030.
NP EN 12390-3 (2009). Testing hardened concrete, Compressive strength
of test specimens, IPQ, Lisbon.
NP EN 197-1 (2001). Cement, Part 1: Composition, specifications and
conformity criteria for common cements, IPQ, Lisbon.
NP EN 206-1 (2007). Concrete, Part 1: Specification, performance,
production and conformity, IPQ, Lisbon.
NT BUILD 492 (1999). Chloride Migration Coefficient from non-
steady state migration experiments, NORTEST.
PE-002 (2005). Concrete, Electrical resistivity - Measuring the electrical
resistivity of a concrete element. Construction Materials Laboratory,
University of Minho, p. 9 (in Portuguese).
Poon, C. S., Kou, S. C., and Lam, L. (2006). “Compressive strength,
chloride diffusivity and pore structure of high performance metakaolin
and silica fume concrete.” Construction & Building Materials, Vol.
20, No. 10, pp. 858-865, DOI: 10.1016/j.conbuildmat.2005.07.001.
Ramezanianpour, A. A. and Bahrami Jovein, H. (2012). “Influence of
metakaolin as supplementary cementing material on strength and
durability of concretes.” Construction & Building Materials, Vol.
30, pp. 470-479, DOI: 10.1016/j.conbuildmat.2011.12.050.
Rashad, A. M. (2013). “Metakaolin as cementitious material: History,
scours, production and composition – A comprehensive overview.”
Construction & Building Materials, Vol. 41, No. 6, pp. 303-318,
DOI: 10.1016/j.conbuildmat.2012.12.001.
Sabir, S. S., Wild, S., and Bai, J. (2001). “Metakaolin and calcined clays
as pozzolans for concrete: A review.” Cement & Concrete Composites,
Vol. 23, No. 6, pp. 441-454, DOI: 10.1016/S0958-9465 (00)00092-5.
Siddique, R. and Klaus, J. (2009). “Influence of metakaolin on the
properties of mortar and concrete: A review.” Applied Clay Science,
Vol. 43, Nos. 3-4, pp. 392-400, DOI: 10.1016/j.clay.2008.11.007.
Srivastava, V., Kumar, R., Agarwal, V. C., and Mehta, P. K. (2012).
“Effect of silica fume and metakaolin combination on concrete.” IJ
Civil & Structural Eng., Vol. 2, No. 3, pp. 893-900, DOI: 10.6088/
ijcser.00202030017.
Trejo, D. and Halmen, C. (2006). Evaluation of Thiele metakaolin for
applications in concrete, Progress Report, Texas A&M University,
p. 36.
Zeljkovic, M. (2009). Metakaolin effects on concrete durability, Master
Thesis. University of Toronto, p. 163.